|Publication number||US6725101 B2|
|Application number||US 09/736,520|
|Publication date||Apr 20, 2004|
|Filing date||Dec 13, 2000|
|Priority date||Dec 13, 2000|
|Also published as||US20020072812|
|Publication number||09736520, 736520, US 6725101 B2, US 6725101B2, US-B2-6725101, US6725101 B2, US6725101B2|
|Inventors||Rene E. Sanchez, Joannes N. M. de Jong, Lloyd A. Williams, Vincent M. Williams, Daniel W. Costanza|
|Original Assignee||Xerox Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (9), Non-Patent Citations (2), Referenced by (9), Classifications (8), Legal Events (6)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
The present invention relates to document processing machines and, more particularly, to document processing machines incorporating control systems.
2. Prior Art
Various document systems require that an automatic document handling apparatus within the system operate within precision tolerances so as to align a particular edge of a document with a given coordinate. For example, in an x-y coordinate system this could be the coordinates 0,0, on a copier platen where the scanner or copier optics have also been aligned to begin imaging at the 0,0 coordinates on the copier platen. This process of aligning the document relative to a given coordinate is known as document registration. In a manual copying operation, an original document is registered by the user with ruler markings along the sides of the platen. In systems having automatic document handling systems document registration is automatic and is generally accomplished by the use of various belts, rollers and sensors situated along the document path. The precision control of the various belts and rollers along the document path is generally accomplished using classical feedback control systems. If the original document is not properly registered then undesirable dark borders and shadow images may appear on the copy. In addition to document registration other document positioning functions requiring feedback control includes photo receptive (PR) drum motion control, and intermediate belt transfer (IBT) motion control, and developer motion control. In general, the desired positioning accuracy for document handling is on the order of one millimeter. As is readily appreciated an automatic document handling system must rapidly and accurately position each document. Conventional feedback control systems correct for the undesired linear motion of the document handling systems. However, the non-linear motion due to motor torque ripple, gear run-out, and roll run-out can not always be corrected due to bandwidth limitations in the classical controller. Currently, the non-linear motion due to these errors is minimized by demanding that tolerances associated with manufactured parts be as small as possible. Achieving such tolerances are difficult and expensive to maintain.
In accordance with one embodiment of the invention, a non-linear control (NLC) system for controlling performance of a device is provided. The NLC system comprises a controller system with a first and second feedback compensator connected to the controller. The first feedback compensator compensates for linear error performance of the device while the second feedback compensator compensates for non-linear error performance of the device.
In accordance with another embodiment, the invention includes a method for controlling non-linear motion of a system. The method comprises the steps of sampling non-linear performance of the system and determining at least one frequency component of the sampled non-linear performance. Next, at least one sinusoid substantially equal in magnitude to the at least one frequency component and substantially 180 degrees phase inverted with respect to the at least one frequency component is applied to the system controller.
Another embodiment of the invention is directed towards a hybrid learning feedback controller (HLC) system for controlling operation of a xerographic imaging device; the xerographic imaging device having an input port and an output port. The HLC feedback controller system comprises a feed forward controller system operative to receive an actuating signal and in response to the actuating signal output a manipulated variable. The feed forward controller comprises a feed forward controller, a first input summing node, a second input summing node, a first proportionality device, and a second feedback compensator having a learning controller system for correcting non-linear motion of the device.
The foregoing aspects and other features of the present invention are explained in the following description, taken in connection with the accompanying drawings, wherein:
FIG. 1 is a perspective view of a conventional document processing apparatus;
FIG. 2 is a schematic view of paper paths and belts within the document processing apparatus shown in FIG. 1;
FIG. 3 is a block diagram of a hybrid learning control system incorporating features of the present invention for controlling belts shown in FIG. 2;
FIG. 4 is a flowchart of one method for learning the phase and amplitude of the sinusoidal signals to be applied to correct for non-linear motion using the hybrid system shown in FIG. 3;
FIG. 5 is a detailed flowchart of one method for finding the phase of each error component frequency shown in FIG. 4;
FIG. 6a is a frequency domain graph showing belt velocity variation before learning and applying the phase and amplitude of the sinusoidal signals as shown in FIG. 4;
FIG. 6b is a frequency domain graph showing belt velocity variation after learning and applying the phase and amplitude of the sinusoidal signals as shown in FIG. 4;
FIG. 7 is a progression graph showing the velocity variation while the algorithm shown in FIG. 4 is finding the minimum velocity variation; and
FIG. 8 is a comparison graph showing belt speed variation before and after learning the phase and amplitude of the sinusoidal signals as shown in FIG. 4.
Although the present invention will be described with reference to the single embodiment shown in the drawings, it should be understood that the present invention can be embodied in many alternate forms of embodiments. Referring to FIG. 1, there is shown a perspective view of a document processing apparatus 10. The apparatus 10 could be any suitable type of document processing apparatus, such as a copier, a facsimile machine, a scanner, a computer printer, or a multifunction device having two or more functions. Referring also to FIG. 2, in this embodiment the apparatus 10 is a copier which includes an original document feed system 11 and a copy document feed system 13. A scanner or image obtainer 12 is provided under a transparent glass platen 15. The scanned information from an original document fed through the original document feed system 13 is imaged onto paper selected from paper trays 14 or 16. Paper selected from either of the paper trays 14,16 is moved by the copy document feed system 13 through the apparatus 10 by means of various intermediate belt transfers 18 and rollers 19 schematically depicted in FIG. 2. The original document feed system 11 also comprises suitable belts and rollers for automatically moving original documents through the document system. Referring now to FIG. 3 there is shown a block diagram of a hybrid learning control system, incorporating features of the present invention, generally intended to be used for controlling document transport systems in a document processing apparatus such as described above with reference to FIGS. 1 and 2 (e.g.: a copier, a facsimile machine, a computer printer, a scanner, or a multifunction device). The hybrid controller consists of a controller 340, a feedback path consisting of feedback gain devices 334, another feedback gain device 341 and feedback transfer element 335; a learning feedback system consisting of a learning feedback element 332 and learning feedback signal generator 331. In addition, the hybrid controller consists of summing nodes 339,338 and amplifier 336.
The controller 340 controls the velocity of IBT (intermediate belt transfer) belts, photoreceptor drums, etc., while the learning controller 332 corrects for the non-linear motion. The combination of the controller 340 with the learning controller 332 provides precise registration between the original image and the surface receiving the reproduced image. Precise registration prevents the appearance of defects (e.g., a border between two colors not present in the original image) caused by slight misalignment of the marking element with respect to the corresponding area of the image receiving surface (e.g., a photoreceptor or a marking medium) at the time of the reproduced image is transferred. The learning controller 332 corrects for the non-linear motion by setting the signal generator 331 to add a sinusoidal signal of equal magnitude and substantially opposite phase for every error frequency to be corrected to summing node 338.
Referring now to FIG. 4 there is shown a flowchart of one method for learning the phase and amplitude of the sinusoidal signals to be applied to correct for non-linear motion of the controlled plant (FIG. 3, item 333). The learning device (FIG. 3, item 332) digitizes S1 the output signal of the controlled plant (FIG. 3, item 333) and determines S2 each of the error component frequencies. The voltage magnitude of the correcting signal is found by dividing the velocity variation of each error component frequency by the transfer function of the velocity/voltage. The transfer function of velocity/voltage vs. frequency can be obtained analytically or experimentally and are determined S3 through well known techniques and is not discussed here. The phase of each error component frequency is determined S4 and is explained in more detail below. Once the phase and amplitude is known the signal generator (FIG. 3, item 331) is set S5 to apply a sinusoidal signal equal in magnitude and substantially opposite in phase for each error component frequency.
Referring now to FIG. 5 there is shown a detailed flowchart of one method for learning (FIG. 4, item 54) the phase of the sinusoidal signals to correct for non-linear motion using the hybrid system shown in FIG. 3. The plant (FIG. 3, item 333) output performance (i.e., belt speed) is discreetly sampled 412 at a constant interval, digitizing and transforming the sample signal from the time domain to the frequency domain. Generally, the sampled error signal is sampled 412 at a rate compatible with a Fast Fourier Transform (FFT) implementation. The phase of the error signal is synchronous to the position of the part that generates the error, this means that the phase of each error signal is referenced from the start of each revolution. Accordingly, to determine the phase of the error signal the start-phase variable is initially set to zero degrees 414 while the limit-phase variable is set to 360 degrees 414. Next, a decision operation determines 415 the start position of the part being controlled and the signal generator (FIG. 3, item 331) applies 416 a sinusoid equal in magnitude at a phase equal to the following:
The plant (FIG. 3, item 333) output performance (i.e., belt speed) is again discreetly sampled 417 at a constant interval, digitizing and transforming the sample signal from the time domain to the frequency domain. The velocity variation amplitude of the frequency of interest is determined 418 and a decision operation 421 compares the velocity variation amplitude of the frequency of interest to a specified error tolerance. If the result of the decision operation 421 is less than the specified error tolerance the phase for the error frequency is set 419 to:
and learning stops 420. If the result of the decision operation 421 is not less than the specified error tolerance a decision operation again determines 422 the start position of the part being controlled; the signal generator (FIG. 3, item 331) applies 425 the sinusoid equal in magnitude at a phase equal to the following:
The plant (FIG. 3, item 333) output performance (i.e., belt speed) is again discreetly sampled 426 at a constant interval, digitizing and transforming the sample signal from the time domain to the frequency domain. The velocity variation amplitude of the frequency of interest is determined 428 and a decision operation 429 compares the velocity variation amplitude of the frequency of interest to a specified error tolerance. If the result of the decision operation 429 is less than the specified error tolerance the phase for the error frequency is set 430 to:
and learning stops 420. If the result of the decision operation 421 is not less than the specified error tolerance a decision operation determines 432 if the velocity variation amplitude determined by step 418 is less than the velocity variation amplitude determined by step 428. An affirmative determination results in setting 432 the start-phase variable to the last known value of the start phase variable and setting 432 the limit-phase variable to:
A negative determination by decision operation 432 results in setting 436 the limit-phase variable to the last known value of the limit-phase variable and setting 436 the start-phase variable to:
The steps are then repeated until a decision operation, 421 or 429, determines the velocity variation amplitude of the frequency of interest is less than a specified error tolerance and learning stops, 420 or 431.
In test, mathematical models simulated gears and roll run-out. In one test the center was set to 0.1 mm off-center, resulting in velocity changes of about 1.814 mm/sec (zero to peak) or 0.565% of ΔV/V. Typical motion quality for applications affecting the image formation in printers requires that %ΔV/V velocity be less than 0.1%. After learning the %ΔV/V was reduced to 0.00291%.
Referring to FIG. 8 there is shown a comparison graph showing the belt speed variation before and after learning the phase and amplitude of the sinusoidal signals according to the steps shown in FIGS. 3 and 4. Referring to FIG. 6a there is shown a frequency graph showing belt velocity variation before learning and applying the phase and amplitude of the sinusoidal signals as shown in FIG. 4. Referring to FIG. 6b there is shown a frequency graph showing belt velocity variation after learning and applying the phase and amplitude of the sinusoidal signals as shown in FIG. 4.
Referring now to FIG. 7 there is shown a progression graph showing the belt speed and velocity variation while the algorithm shown in FIG. 4 is finding the minimum velocity variation. It is readily appreciated from FIG. 7 that learning in a particular system is done once, unless one of the parts is replaced. For example, if a motor is replaced, the magnitude and phase of the correction signal should be learned again. It is also readily appreciated from the descriptions above that the disadvantages of the prior art are overcome by the present invention. Specifically, the invention allows manufacturing tolerance requirements such as pinion run-out tolerances to be relaxed by dynamically compensating non-linear motion induced by finite tolerance.
It should be understood that the foregoing description is only illustrative of the invention. Various alternatives and modifications can be devised by those skilled in the art without departing from the invention. Accordingly, the present invention is intended to embrace all such alternatives, modifications and variances which fall within the scope of the appended claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3652942||Jan 18, 1971||Mar 28, 1972||Lummus Co||Feedback control system|
|US4490841 *||Oct 21, 1982||Dec 25, 1984||Sound Attenuators Limited||Method and apparatus for cancelling vibrations|
|US4547858 *||Jun 13, 1983||Oct 15, 1985||Allied Corporation||Dynamic control for manipulator|
|US4727303||May 22, 1986||Feb 23, 1988||Gmf Robotics Corporation||Positional control method and system utilizing same|
|US5124626||Dec 20, 1990||Jun 23, 1992||Mts Systems Corporation||Sinusoidal signal amplitude and phase control for an adaptive feedback control system|
|US5491594 *||May 20, 1993||Feb 13, 1996||Sharp Kabushiki Kaisha||Magnetic recording/reproducing apparatus|
|US5605097 *||Apr 26, 1995||Feb 25, 1997||Schablonentechnik Kufstein Aktiengesellschaft||Method for producing a screen printing stencil with a laser|
|US5950040||May 22, 1998||Sep 7, 1999||Xerox Corporation||Feedback control system for controlling developability of a xerographic imaging device|
|US6326758 *||Dec 15, 1999||Dec 4, 2001||Reliance Electric Technologies, Llc||Integrated diagnostics and control systems|
|1||*||Breitfelder, Kim, and Messina, Don, IEEE 100: The Authoritative Dictionary of IEEE Standards Terms, Standards Information Network IEEE Press, 7th Ed., 879.|
|2||*||Pham, Duc Truong & Liu, Xing, Neural Networks for Identification, Prediction and Control, 1995, Springer-Verlag London Limited, 1st Ed., 175-76.*|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7103426 *||Jan 28, 2003||Sep 5, 2006||Rosemount Analytical Inc.||Anticipatory high frequency noise compensation in a distributed process control system|
|US7157873||May 2, 2005||Jan 2, 2007||Xerox Corporation||Systems and methods for reducing torque disturbance in devices having an endless belt|
|US8340477 *||Mar 31, 2008||Dec 25, 2012||Intel Corporation||Device with automatic image capture|
|US8622938 *||Aug 28, 2009||Jan 7, 2014||University Of Tsukuba||Wearable action-assist device, and method and program for controlling wearable action-assist device|
|US8649885||Nov 25, 2008||Feb 11, 2014||Nikon Corporation||Frequency selective iterative learning control system and method for controlling errors in stage movement|
|US20040148038 *||Jan 28, 2003||Jul 29, 2004||Keyes Marion A.||Anticipatory high frequency noise compensation in a distributed process control system|
|US20060244404 *||May 2, 2005||Nov 2, 2006||Xerox Corporation||Systems and methods for reducing torque disturbance in devices having an endless belt|
|US20090245695 *||Mar 31, 2008||Oct 1, 2009||Ben Foss||Device with automatic image capture|
|US20090319054 *||Dec 24, 2009||University Of Tsukuba||Wearable action-assist device, and method and program for controlling wearable action-assist device|
|U.S. Classification||700/45, 700/261|
|International Classification||G05B13/02, G05B21/02|
|Cooperative Classification||G05B21/02, G05B13/0265|
|European Classification||G05B21/02, G05B13/02C|
|Dec 13, 2000||AS||Assignment|
|Dec 18, 2000||AS||Assignment|
|Jul 30, 2002||AS||Assignment|
Owner name: BANK ONE, NA, AS ADMINISTRATIVE AGENT,ILLINOIS
Free format text: SECURITY AGREEMENT;ASSIGNOR:XEROX CORPORATION;REEL/FRAME:013111/0001
Effective date: 20020621
|Oct 31, 2003||AS||Assignment|
Owner name: JPMORGAN CHASE BANK, AS COLLATERAL AGENT,TEXAS
Free format text: SECURITY AGREEMENT;ASSIGNOR:XEROX CORPORATION;REEL/FRAME:015134/0476
Effective date: 20030625
|Aug 24, 2007||FPAY||Fee payment|
Year of fee payment: 4
|Aug 12, 2011||FPAY||Fee payment|
Year of fee payment: 8